Posted by: @ian33a

↑

With an 18KW heat loss calculation and concerns about installers and poor SCOP achieved, I wonder if riding out Trump's warmongering and hoping that the cost of oil will drop is a better proposition. That said, if I do the maths using a formula that I found on another thread on this site, with our oil usage over the past 12 months, being about 2010 litres, we would need a heat pump of around 8KW.  

I do appreciate that an accurate figure only comes from a proper heat loss survey and this, quite rightly, involves time, effort and the use of software which needs to be paid for. I wonder though, why the on-line estimate is twice as much as what appears to be real use data - are our delta T's different, are incorrect assumptions from an out of data EPC being used or is it something else?

2010litres is indeed likely to be about 8kW,  although you need to add perhaps 20% for 18C vs 20C which is the normal design temperature, still nowhere near 18kW. 

Its not uncommon for surveyed results in older properties that have been upgraded to be way out.  My own 1930s (originally solid wall) house with reasonable but not extreme fabric upgrades is 7kW measured, the 'full' surveys said 16kW, if I use the MCS assumptions I can get to 10.5kW, but only if I use an ACH of 0.5 can I get to 7kW.  I have a 7kW heat pump and its just right.

In your circumstance I would not proceed without getting to the bottom of this.  You could do your own heat loss calculation which would ensure that the surveys dont overlook fabric improvements and then play with ACH values, or ghet a professional survey done but check the assumptions. It sounds to me like you may well need a 10-12kW device which, unlike 18kW, is readily obtainable.  With 18kW you may be into cascades, you dont want to go there unless there really is no alternative.

With gas its possible to do a more sophisticated analysis - see mine attached - not so easy with oil unless your boiler produces stats or you have metering

Hi @ian33a you could do your own heat loss calculation iterating with the conductivity of your walls until you get to the less than 10kW you probably need. 
I’m not surprised that a rough ‘type of house and size’ calculation shows such a high power needed since your walls would be much lower loss than the classic no cavity brick.

I found Heatpunk very good https://heatpunk.co.uk  

A measurement of your actual air leakage will cost you about £300 and that is a valuable input into both your calculations and a potential installer.

Thank you both for your replies. It is heartening to discover that our house is no-where near as bad as the Heat Geek prediction algorithm suggests it would be.

I did start down the route of doing my own heat loss calculations but didn't pursue it too far as we have a house which is very much hybrid - about a third cob, a third thick stone and a third block cavity. Some of the walls have internal dry lining, some don't. Some of the roof is accessible and the insulation deterministic, other parts are vaulted and I have to make assumptions about what insulation there is in the voids. What insulation is behind the drylining isn't clear.

With this variety and soup of unknowns was why I reverted to considering actual oil usage over a period as an indicator.

Thermal imaging on a regular basis suggests that our house isn't riddled with loads of holes although some of the exposed stone walls are, obviously, colder than the insulated cavity ones. 

I had noticed heat punk before but hadn't gone too far into it. I'll have another play. It certainly appeals to my engineering background!

Posted by: @ian33a

↑

With this variety and soup of unknowns was why I reverted to considering actual oil usage over a period as an indicator.

Using past oil usage to estimate heat loss is tricky.

Firstly it is very unusual to have accurate interval usage. You fills your tank and runs your heating, but how much oil did you actually use between two dates? To estimate heat loss from past fossil fuel usage needs daily usage, which you then plot against daily energy delivered to the property, which in turn needs an estimate of the boiler efficiency, which is at best an estimate, and very probably a guesstimate. Past annual oil use times a 'magic number' may or may not provide a sanity check.  

Secondly, if you ran your oil boiler on a timer, eg morning and evening, your use pattern will change when you have a heat pump. It will run all of the time, or most of the time, and as a result you will almost certainly use more heat (energy).

Heat loss calculations done using a spreadsheet are just standard whatiffery. What if my walls are X, my windows are y etc. If you do one yourself, and get the 'what ifs' right, it won't be miles out. Hybrid construction isn't a problem, just use the right U values for the different materials. You may need to do some research on them, some standard spreadsheet defaults can be out eg solid stone walls may not be quite as leaky as some would have you believe. As others have said, you also need to use reasonable ACH (air change per hour) estimates.

I suspect many commercial installers deliberately bias their spreadsheet calculations to ensure an over-size heat pump recommendation. This is great for them because it reduces the risk of supplying an under-powered heat pump, but it is not such good news for consumers. They get a bigger heat pump than they need (more £££), and at higher OATs, the heat pump is likely to struggle to lower it's output sufficiently, resulting in cycling. That said, cycling isn't always a bad thing. Slow cycling (once an hour or so) to modulate output when demand is low doesn't appear to have any drawbacks. 

@cathoderay Very valid points and I agree with your statements.

FWIW, and it's far from ideal, I measure oil usage at the start of every month so I have 12 points of reference. Hardly 365 and it doesn't take into account outside temperature (although I should start to have that kind of data having installed an outside heat sensor linked to Home Assistant which I have fitted recently).

Our oil boiler drives a thermal store and from this we get DHW as well as heating. The heating is via hydronic under floor with all zones (10 of them) running with the same thermostat profile (18 degrees by day and a 2 degree set back over night). The boiler is active 24/7 and cuts in as and when the tank stat goes below a certain threshold. I also bypass the boiler automatically if electricity is cheap or if we have an over abundance of power in our batteries (using HA) but this is a summer event mostly and we turn off the heating completely during the summer months as the thickness of the house walls act as a massive thermal heatsink to keep the house temperature decent enough.

I'm confident that I am measuring oil usage reasonably accurately as I check the levels with a dip stick and calculate the remaining volume based upon the shape of the tank. It lines up very accurately when a tanker refills all or part of it. Not as good as a meter, but decent enough.

But yes, a proper heat loss calculation is best. I plan to have a go with one of the free packages, perhaps with a bit of online training and just see what comes out. The hardest part will be estimating the heat loss through the walls - some of which are 300 years old  and with a make up that I can only guess!

@ian33a — a measured heat loss always trumps a calculated (whatiffery spreadsheet) heat loss, providing certain requirements are met.

Firstly, the IAT has to be stable, because this is basis of the method, if the IAT is stable, then the heat loss equals the heat being put into the house. This means no timers, not setbacks etc during the assessment, just a stable constant IAT.

Secondly, you need to have daily energy supplied to the house data. If you can measure your monthly oil consumption, could you also on occasion measure the daily consumption to reasonable accuracy? I suppose it will depend very much on the shape of your tank, and so by how much the vertical level as measured by a dip stick changes day by day. If you can, and you can also make an educated guess at your boiler's efficiency, then you have daily energy supplied to the house data.

I think there may well be other ways of measuring the level which might be more accurate. Presumably the automatic monitoring and top up when required systems have some form of monitoring (ultrasound?) - might that be an option?

Thirdly, you need the mean OAT. You already have a way of getting that.

Given the above, I would collect this daily data on days with a range of OATs and then plot daily energy delivered to the house against daily mean OAT. I don't think you will need a lot of data points, just a good range of OATs, and if they all line up then you probably have a valid result.

Of course this requires data collection during the heating season, which we are now leaving, and the next one won't come around again until next autumn, which may or may not fit with your proposed installation timescale.

@cathoderay Yes, a very fair analysis once again and agreed with.

Our tank is 2000 litres or so and, when full, the oil is about 955 mm deep. Being cylindrical, the rate of use vs the dipstick reading won't be linear and will be especially hard to calculate when the tank is about half empty - especially with the precision of a tape measure lined up against a length of slightly rusty steel bar!

At the moment 1mm of oil equates to about two litres of oil and our daily usage at the moment is around 2 litres per day. Admittedly though, when it's really cold in the winter, the daily usage can be six times that and is more measurable, albeit with a need for some averaging given the resolution of measurement and human eye interpolation with a tape measure!

We have an ultrasonic device for measuring in the tank but the internal display is graduated at 10% increments - and so is useless as an accurate measuring device. 

I tried a professional laser measuring device but the measurement was totally wrong, but consistent - so no use!

Your suggestion of daily measurements does make sense (when the usage is higher), perhaps with some margin of error built in to address the limitation of dipstick measuring. There's no massive hurry for us to engage with a heat pump installer - unless Trump becomes even more unhinged and oil hits and stays at $200 a barrel !

Just as a side note - a friend of mine had a heat pump installed about three years ago. He has a 5 bedroomed house, constructed within the last 10 years, timber framed, brick exterior, and very much with efficient running costs in mind. All underfloor hydronically heated. About 280 square metres, detached and on three floors.

His heat pump is 16KW !!!

I wonder what the surveyor was smoking when he did the survey on that house? Interestingly, my friend only chooses to heat the ground floor and just when Octopus Cosy is cheap, otherwise the heat pump is turned off - that sounds like a very inefficient and poorly specified system to me  .... but maybe I am wrong?

@ian33a — I have a sort of residual interest in this problem as I used to have an oil tank (and boiler) and I remember how vague my usage measurements were. I had a sight gauge, and with a 4 x 2ft tank (8sq ft, 0.74 sq m), a 1mm drop represented 0.74 litres. But how to measure the level to the nearest mm? Hooking a tape measure on the bottom of the tank and eyeballing the level probably had enough parallax error to mean the individual readings were somewhat dubious. I also know about cylindical tanks and dipsticks, I have that setup on a boat. The dip stick is marked off in quarters...however, I do know how much fuel I use on average, by recording refill to refill volumes, and noting the engine hours on the engine hour meter. Which leads on to...  

Doing a non-AI assisted google search, I am reminded that oil burner nozzles are calibrated (and so marked) to deliver a certain flow rate. Some clever folk have used that plus a simple boiler pump motor on/off data logger (which can even be DIY with a bit of ingenuity) all hooked up to something like Home Assistant to measure their oil use. They say it is a simple reliable method and accurate enough (though they don't give actual figures apart from within a few gallons on refilling). One enterprising punter even disabled the ignition, ran the boiler and used a bucket to measure flow rate from the nozzle. See this openenergymonitor page: https://community.openenergymonitor.org/t/measuring-oil-flow-to-home-heating-boiler/11384.

If I still had a oil fired boiler, I would definitely start exploring these ideas.

Posted by: @ian33a

↑

His heat pump is 16KW !!!

Good grief! He could try using the whole house as a slow cooker! But the other idea, banking some heat when tariffs are low has come up on the forum before. I don't have a ToU tariff, but others do, and they may care to comment. @jamespa may have some useful thoughts as well.  

Not sure how heat geek ‘estimate’ heat loss online but if based on an EPC then in an old house of hybrid construction including cob then likely to be complete fiction which is probably why they don’t want to touch it. We were in a similar situation - Estimate based on EPC required 20kw heatpump in a stone house- did a thing where I measured oil boiler burner run time during a cold snap and steady state conditions and estimated 14kw. Heatpump installer suggested a 14kw Ecodan. Have since fitted a heat meter  and open energy monitor and actual measured heat loss nearer 10kw so half where we started - but we do have an AGA which will add maybe 2kw?? So not far out in the end.

If you have the time and inclination fit a heat meter with a basic open energy monitor setup - my heat meter was about £300?? (plus install but I DIY) if I remember correctly plus a Rasberry pi for £40 and open energy monitor software (free) , add and inside and outside temp sensor and you will get a far more accurate heat loss than any heat loss survey. You may have trouble getting an installer to believe you though. You can then re use the heat meter with your new heat pump and become a sad person like me😵‍💫

@travellingwave That's really interesting information, thank you.

Please could you point me toward the heat meter that you installed (or one similar). Ideally, I'd install one which works with Home Assistant but wouldn't be against installing one with is OEM compliant (I've worked with both OS platforms). 

As a retired engineer, I'm pretty practical and probably a I'm a sad individual too!

Just checked and a DN25 heat meter plus fittings is £325 from OEM plus you need an Mbus dongle (£46)

To install would require draining down all or part of your system and cutting into flow and return pipes to boiler to fit meter. Probably half a days work for a good plumber or fairly DIYable if confident with plumbing.

Another Much cheaper though less accurate option - get a Rasberry pi - install open energy monitor - add temp sensors to measure IAT , OAT and flow a return temps. You then just need water flow rate in order to calculate heat delivered. We have a Wilo pico circulating pump on our system - this pump does have a display which amongst other things gives an estimate of flow rate - and in fact comparing it to the flow rate measured by my heat meter is pretty accurate. Assuming you don’t have lots of zone valves and TRVs switching on and off you could use this value to calculate heat delivered based on your flow and return temps. If you don’t have a modern circulation pump they are about £130 and fairly easy to swap out as there should be isolation valves either side of the pump.

Apologies if I have gone on a tangent- however given the “challenging” construction of your property a bit of time spent up front May be worthwhile.